High-quality　fire　image　is　essential　for　fire　detection　and　monitoring.　Degraded　fire　images　in　the　fire　field　aggravate　the　difficulties　in　determining　the　burning　area　and　assessing　the　severity　of　the　fire,　thus　being　unable　to　make　scientific　firefighting　strategies.　Therefore,　a　fire　image　enhancement　model　of　multi-segment　resolution　structure,　designed　to　implement　image　enhancement　step　by　step,　is　proposed　to　solve　the　problem　of　low-resolution　and　detail　reconstruction.　Through　testing　samples　by　both　the　experimental　and　simulated　datasets,　the　experimental　results　show　that　the　model　significantly　improved　image　quality,　with　the　peak　signal-to-noise　ratio　increasing　by　140.66%　in　the　experimental　dataset　and　by　3242.67%　in　the　simulated　dataset,　while　the　structural　similarity　index　measurement　increased　by　525.90%　and　3090.07%,　respectively.　These　enhancements　led　to　an　85%　improvement　in　the　recognition　accuracy　of　the　fire　detection　model　in　the　experimental　dataset　and　a　55%　improvement　in　the　simulated　dataset.　The　model　also　exhibited　strong　robustness,　effectively　restoring　flame　contours　in　images　with　varying　smoke　concentrations　and　fire　sizes,　suggesting　its　potential　for　application　in　complex　fire　scenarios.　The　proposed　method　illustrates　the　effectiveness　of　the　multi-segment　resolution　structure　in　enhancing　fire　images,　providing　a　novel　approach　to　improving　fire　monitoring　quality.　The　fire　image　enhancement　method　based　on　artificial　intelligence　has　far-reaching　significance　for　advancing　the　informatization　and　intelligence　of　fire　protection.